Search Results for "malware-samples"

...The idea is to distinguish between fricatives and voiced sections of the speech signal by the number of zero crossings in time. Most of the speech file is left untouched (the samples are directly copied from source to destination). Only fricatives that are long enough and loud enough are filtered. The advantage of this approach over traditional approaches is that the clarity of the remaining speech is completely unaffected.